Mind your Context
Feb 2, 2026 • 3 min read
I was losing 47% of my context window to overhead before Claude did any real work. Here's how I dropped that to under 1%.
Feb 2, 2026 • 3 min read
I was losing 47% of my context window to overhead before Claude did any real work. Here's how I dropped that to under 1%.
Jan 27, 2026 • 3 min read
Claude Code's native sub-agents can't compact their own context. Here's how I built true agent swarming using headless instances and an orchestrator pattern.
Jan 16, 2026 • 3 min read
Here's the uncomfortable truth: AI coding tools are fluent in logic but illiterate in experience. You can't feel the interface until the code is done. You've burned three days on backend architecture, database schemas, API endpoints, and deployment configs. Then you finally see the UI. It's wrong. The user flow is awkward. The data you're surfacing isn't what users actually need. The interaction patterns don't match how people actually think. Now you're refactoring everything. Again. Twenty-t
Jan 16, 2026 • 5 min read
I've built a spec-driven workflow using Claude Slash Commands that mirror the best SDLC processes.
Nov 8, 2025 • 6 min read
Let’s Talk About the Elephant in the Room Here’s something nobody wants to admit: AI coding assistants are making most developers slower. Not a little slower. 19% slower. And here’s the kicker — those same developers believed they were 20% faster. They expected the AI to speed them up by 24%. Meanwhile, the screen recordings and time logs told a different story entirely. If you’re using GitHub Copilot, Cursor, Claude Code, or any AI coding assistant, you need to hear this. Because there’s a
Nov 8, 2025 • 13 min read
In Article 1, we talked about the uncomfortable truth: AI coding assistants are making most developers 19% slower, not faster. We covered the six problems: 1. The Performance Trap (verification overhead) 2. The Code Quality Death Spiral 3. The Attention Problem (information overload) 4. The Context Problem (AI doesn’t know your project) 5. The Trust Problem (hallucinations and false confidence) 6. The Security Problem (45% vulnerability rate) If you did the homework — tracking one codin